Analyzing and Responding to Repeat Offending
Tool Guide No. 11 (2016)
by Nick Tilley
The Problem of Repeat Offending
A wide range of research converges on the following findings about criminal offenders:
- Some level of participation in criminal activity is normal, especially during adolescence and among males.1 Almost all citizens act dishonestly, commit crimes, and behave in antisocial ways at some point in their lives.† Most will have committed more than one crime.2
- Most people offend infrequently and soon age out of committing crime. Involvement in criminal behavior peaks in adolescence (ages 14–17) and then generally fades rapidly.3
- A much smaller number of persistent and prolific offenders are responsible for a substantial proportion of all crime.4 Roughly half the crimes committed can be attributed to those identified as prolific offenders. Males commit far more offenses than females do, but even among female offenders, a small percentage commits a hugely disproportionate number of the offenses.
That a small fraction of offenders commits a large fraction of crime may come as no surprise to most police officers. The breadth of low-levels of offending and the proportion of crime attributable to those involved in it may not be so widely understood.‡
‡ For a thorough review of these patterns, see Blumstein et al. (1986). They summarize findings as follows: “The median offender commits only a handful of crimes per year, while a small percentage commit more than 100 crimes per year.” (p. 4)
There are two general theories of repeat offending patterns. One theory is that some people are highly disposed to behave criminally, and this leads them to sustained criminal careers in which they offend frequently. These “lifetime persistent” offenders begin offending early and have long crime careers. They are distinguished from “adolescent limited” offenders, who start later and finish earlier, as the name suggests.5 Another theory suggests that one criminal act begets another. That is, involvement in one crime increases the probability of further offending. For example, someone convicted of a crime finds it more difficult to resume a law-abiding life, either because they have fewer job opportunities or because they are shunned by normally law-abiding members of the community. Therefore, they persist in criminal behavior and associate with others who are in a similar position. It might also be that the rewards of successfully committing crime reinforce the criminal behavior and make persistent offending more likely.
Both theories play a part in producing repeat offending patterns. Much has been written on efforts to identify and deal more effectively with individuals who have strong dispositions to commit crime and to avoid inadvertently precipitating high-rate offending among those who might otherwise not be drawn into criminal careers. The main focus of this guide, however, is how to identify problems that call for special police attention to prolific offenders and how to devise, implement, and evaluate effective strategies to deal with them.
If police can identify prolific offenders early enough and implement effective measures to reduce their rate of offending, crime levels can be reduced in most instances.
There is a risk that multiple contacts with the criminal justice system unintentionally provide “stepping stones” toward further criminal involvement, especially among the poor.6 Exclusionary processes can reduce opportunities to follow a law-abiding life. Labeling can alter an individual’s identity. These processes can interlink, as in the following sequence:
1. An individual is disadvantaged, with few life chances.
2. He commits crimes, as is very common among adolescent males.
3. The police arrest him.
4. He does not have friends and family capable of credibly speaking on his behalf.
5. He is prosecuted and convicted.
6. His education is disrupted.
7. He is also stigmatized, resulting in fewer employment opportunities and a smaller network of contacts that might help him find work.
8. His criminal record and lack of employment inhibit him from associating with people who might have a positive influence on him.
9. As his opportunities for a law-abiding life diminish and his integration with law-abiding members of the community atrophies, he spends more time with other offenders.
10. He identifies as a criminal and absorbs a rhetoric from other offenders that legitimizes his criminal behavior.
Although common, this pathway is far from universal: circumstances (or turning points) that provide branches pointing in different directions often arise (or are created).7 Therefore, when devising strategies to deal with repeat offending problems, you need to consider whether your agency inadvertently contributed to their production, and explore partnering with others who are better placed to help offenders become law-abiding citizens by either exerting informal control over them, providing them with legitimate opportunities, and/or creating turning points away from crime.
The Relationship of Repeat Offending to Repeat Victimization and Repeat Locations and Targets
Crime is disproportionately concentrated not only on a small subset of offenders but also small subsets of victims, targets, and places. Patterns of repeat victimization—whereby experiencing one crime increases the probability of experiencing another—have been found across diverse crime types and in diverse places.† Likewise, crime is concentrated on hotpots,‡ and, in the case of theft, on so-called “hot products.”§
† See Problem-Solving Tool Guide No. 4, Analyzing Repeat Victimization, for further information.
‡ See Problem-Solving Tool Guide No. 6, Understanding Risky Facilities, for further information.
§ See Problem-Solving Tool Guide No. 12, Understanding Theft of ‘Hot Products’ for further information.
Concentrated repeat offending occurs in some disadvantaged neighborhoods due to the following process:8
1. Inner-city neighborhoods, historically populated by a large black underclass, continue to draw in more young, poor black people while the better-off residents migrate to the suburbs.
2. This departure means the remaining residents lose contacts to help them find work.
3. Changes in the economy shrink the supply of unskilled manual jobs for men.
4. Joblessness as a way of life sets in, and work-related norms wither.
5. Teachers become frustrated by the lack of interest in education, so fail to teach children.
6. There are few men fit to marry so vulnerable families are headed by single mothers.
7. Good, employed male role models are scarce.
8. Segregated, ghettoized public-housing projects emerge, with little reciprocal guardianship or control, thereby creating vulnerable residents.
9. In these neighborhoods, relatively uncontrolled prolific offenders meet relatively unprotected targets. Repeat offending, repeat victimization, and hot spots are therefore concentrated.
The police are clearly in no position to control the wider social structural processes associated with such dynamics. They are mentioned here only as background to be considered and to indicate likely limits to what the police can do.
The police are more likely to be able to address specific problems in specific places where repeat offending may constitute part of the problem and reducing it part of the solution. For some problems there are specific links between hotspots, repeat victims, and prolific offenders. In regard to domestic burglary, for example, high-crime areas experience high rates of crime due in part to the concentration of repeat victims there.9 Repeat offenders are often responsible for those who are repeatedly burglarized, and those offenders tend to be prolific.10 The same pattern has been found for bank robbery and bank robbers.11 A more obvious fact is that domestic violence tends to be a problem of repeat offenders and repeat victims, generally in the same location. Feuds between gangs or neighbors also tend to be problems where the same offenders, victims, and locations are repeated.
For other specific problems, there may be a difference in the populations of repeat offenders, repeat victims or targets, and repeat locations, especially in locations such as shopping malls, which provide rich targets for criminals.† There do appear, nevertheless, to be good reasons to expect that repeat offenders will return to the same or similar targets and places because of their familiarity with the risks and rewards they will face and the successful methods they used in the offenses they committed previously.12
† See Brantingham and Brantingham (1995), who distinguish between a) crime attractors, which are places that attract offenders (some of whom may return repeatedly) because of the rich pickings and b) crime generators, which are also rich in opportunities but have a changing population there primarily for reasons other than to commit crimes, some of whom might be tempted to offend. Shopping malls likely function as both crime attractors and generators.
Predicting Repeat Offending
In an attempt to predict who is liable to become a persistent and prolific repeat offender, many risk factors have been identified.13
- Family-related risk factors include poor supervision and discipline, family conflict, family history of problem behavior, parental involvement in and attitudes condoning problem behavior, and low income and poor housing.
- School-related risk factors include low achievement beginning in elementary school; aggressive behavior, including bullying; lack of commitment, including truancy; and school disorganization.
- Community-related risk factors include a disadvantaged neighborhood, community disorganization and neglect, availability of drugs, high population turnover, lack of neighborhood attachment, and low collective efficacy (social cohesion and informal social control).14
- Individual and peer-related risk factors include alienation and lack of social commitment, attitudes that condone criminal behavior, early involvement in criminal behavior, having friends involved in problem behavior, low intelligence, anxiety, and social awkwardness.
Risk factors have proved imprecise in predicting who will eventually become a prolific offender. Data from what is perhaps the most thorough study of males to date indicate that a third of those who turned out be to persistent offenders showed no evidence of any risk factors at ages 8 and 9. These comprise false negative predictions. Moreover, half of those males with the maximum number of risk factors considered did not turn out to be persistent offenders. They comprise false positive predictions.15 In addition, other research suggests that particular events and experiences, such as moving home, changing schools, joining the military, or falling in love, may serve as turning points that divert individuals toward or away from careers as repeat offenders.16 At a population level, risk factors may be useful in identifying sub-groups whose members are more likely to become repeat offenders. However, at the individual level, these factors have been less successful in pinpointing—at an early age—those who can be expected to later commit many crimes.
Once criminal careers are established and offenders are processed by the criminal justice system, recidivism rates become very high: up to two-thirds of those who are incarcerated will reoffend within a few years.17
Measuring Repeat Offending
Mounting a sophisticated, longitudinal research study to analyze your local repeat offending problem will take too long and be too costly. If you suspect that repeat offending contributes significantly to the problem you are addressing and that it can be reduced, you may wish to make a “good enough” estimate of the contribution made by repeat offenders. For this you will have to depend on readily available sources. The most obvious sources relate to offenders who have been officially processed, either through convictions or through arrests.
Reoffending rates measured through re-conviction are, of course, likely to underestimate real reoffending rates. Steep attrition patterns, going from the number of offenses down to the number for which there is a conviction, mean that the vast majority of offenses do not lead to convictions. They are unreported, undetected, or, for various reasons, not prosecuted or unsuccessfully prosecuted. In the United States in 2009, the clearance rates for reported crime (crimes cleared by arrest or ‘exceptional means,’ for example where the suspect dies or the victim refuses to cooperate) varied from 45 percent for violent crime, down to 12 percent for burglary and 11 percent for motor vehicle theft.18 Sanction detection rates (where the offender receives a formal sanction) for reported
crime in England and Wales are likewise higher for violent crime (44 percent in 2010–11) than for property crime where there is usually no direct contact between offender and victim (for example, 13 percent for burglary, 9 percent for theft from a motor vehicle).19 If there were a consistent detection rate of 20 percent, for instance, then the chance of two successive offenses being detected would be 4 percent (0.2 x 0.2 = 0.04). The chance that a third would also be detected is less than 1 percent. Simply put, many of the offenses committed by repeat offenders do not feature in official police statistics.
There are two ways in which one conviction may influence the likelihood of another. The first is that police pay special attention to those who are arrested, thereby increasing their risk of apprehension. The second is that as offenders commit more crime, they gain more experience and expertise and learn how better to evade the police. So, oddly enough, offenders may become both more and less likely to get caught over time. The best evidence is that most offenders seem to get better at avoiding capture rather than police getting better at catching them, at least in the absence of special efforts to target prolific offenders.20
Thus, any local analysis using your own police data on arrests or convictions is more likely to underestimate rather than overestimate the contribution of repeat offending to your problem, although this may not be the case in specific places and for specific offenses. Your data on repeat offenders/arrestees’ contribution to the total number of offenses detected, or for which arrests are made, will give a rough idea of the contribution of repeat offending to your problem, but you need to consider the findings cautiously and with an eye to local knowledge and prevailing policing practices.
Table 1 on page 14 shows the form of a simple repeat offending analysis you could undertake for your target crime. It shows that 10 offenders were arrested for the crime. They were arrested for 30 offenses in all, 20 of which were related to the local problem being addressed. They were each arrested for an average of two target offenses, although two were arrested for four target offenses. The table shows that of the 10 who were arrested at least once for the target offense, half went on to commit at least two offenses. Of that five, three (60 percent) went on to commit at least three offenses and of these three, two (67 percent) went on to commit four. Therefore, the chances of rearrest increase with the number of previous arrests. The table also shows that only two of the offenders were arrested for only one offense in all, and one-third of the total arrests for the target offenses were for other offenses. The final column shows each arrestee’s current age. A majority
are under 20-years-old and all but one is younger than 30. The youngest and oldest are infrequent offenders. Those in the middle age range appear to account for the more prolific offenders, but the numbers are too small to say anything about this with confidence.
Table 1: Offenders and their repeat offenses
Arrests all crimes
Arrests target crimes
More complex analysis of a data set of these offenses and their known offenders could go on to look at the spatial and temporal proximity of repeat offenses: were they committed in quick succession and were they committed close to one another? You can use the free-to-download Near Repeat Calculator to carry out analysis of this sort.21 Gradually you can build a picture of the known repeat offending patterns, always acknowledging that most offenses of most types remain undetected. It may well be that the 20 target offenses shown in Table 1 represent only 15 percent of all reported target crimes. It might be reasonable to suppose that their apparent contribution to the total number of target crimes is underestimated and that targeting repeat offenders successfully could produce a larger impact than the figures suggest. Moreover, if the measures put in place to deal with prolific offenders for the target crime succeed in inhibiting their criminal behavior in general, this would lead to a reduction in a broader array of offenses.
In all cases, you need to be clear about the problem fueling your interest in repeat offending and the sorts of intervention that could realistically reduce it (which will be covered below). You also need to distinguish varying reoffender types relevant to your problem and for whom you need different strategies. For example, if your problem relates to serious offenses committed by professional criminals, you won’t be interested in, for example, street people who commit a lot of petty offenses, but are unlikely to receive special police attention among a group of more serious offenders. On the other hand, if your problem is nuisance offending, you will be uninterested in career organized offenders. Both of these problems will differ from those of solitary, serial sex offenders, as an example.
Another category might be repeat mental health-service consumers, individuals whose aberrant behavior is caused by mental illness and who might have frequent contact with the police, but who end up being referred for mental-health treatment more often than to the criminal courts. Ex-prisoners may comprise another source of your local problem of repeat offending, as may substance abusers, who are liable to be generalist reoffenders. Repeat traffic-offense violators might comprise yet another category. Not only should the responses to these types of offenders vary, but so too should the eligibility criteria for special attention. Tailor your analysis to the problem at hand. You would need to add further columns capturing relevant attributes to Table 1 to reflect the particular attributes of the problem.
You can often usefully complement the statistical analysis of repeat offending relevant to your problem through qualitative information gleaned from interviewing known offenders and by drawing on intelligence gathered from the community, informants, and local service providers. In all cases, however, it is important to be mindful of “confirmation biases,” which often lead us to attach undue significance to information that reinforces our existing views and to ignore information that contradicts them.